OpenAI CLIP: PyTorch Implementation with BERT Text Encoder

A PyTorch-based implementation of the OpenAI CLIP architecture for image-text alignment, authored by Moein Shariatnia and updated in October 2025. It provides a dual-encoder framework for processing image-text pairs using BERT for natural language processing and Vision Transformer components.

Use Cases

• Training image-text alignment models using the provided PyTorch training scripts
• Extracting semantic features from text using the integrated BERT encoder
• Implementing zero-shot classification by comparing image features against text-based labels

Strengths

• MIT License allows for unrestricted commercial and academic use
• Uses BERT for text encoding, providing high-quality natural language processing
• Built on PyTorch for compatibility with modern deep learning ecosystems

Limitations

• No raw data records or pre-processed datasets are included in the repository
• Requires external sourcing of image-caption pairs for training
• Lack of pre-trained weights necessitates significant computational resources for initial training

Provenance

Source

GitHub repository by moein-shariatnia

Freshness

Last updated October 2025.